No, this Danish study does not prove that the mask is useless

Daily life in Marseille with the compulsory wearing of a mask.

-

• Danmask-19, a Danish study published in the

  Annals of Internal Medicine on

   Wednesday, November 18, is the subject of numerous viral social media posts.

• Researchers tried to understand whether wearing the surgical mask outside the home reduced the risk of Covid-19 infection.

  According to their results, 1.8% of mask wearers contracted the coronavirus, compared to 2.1% of those who did not.

  Proof of its uselessness, for many Internet users.

  For the authors of this study, these results, very partial, are "inconclusive".

• 20 Minutes

   takes stock with Professor Didier Lepelletier, co-chair of the permanent working group on Covid-19 within the High Council for Public Health.

  According to him, the study "has too much bias to provide significant results"

“Stop obscurantism, long live science!

"In a Facebook post, Florian Philippot, leader of the Patriots, praised the merits of a Danish study which would demonstrate" the total uselessness of the mask outside against the Covid-19 ".

"All the truths are coming to light," exclaimed the former right-hand man of Marine Le Pen about this study, which would question the French doctrine on the wearing of the mask against the coronavirus.

In a Facebook post, the leader of the Patriots, Florian Philippot, shares a study that would demonstrate - Tom Hollmann

Similar publications have spread like wildfire on social networks, building on the publication in the journal

Wednesday of the results of a research project dubbed "Danmask-19".

This user, for example, tweeted about the study: “It covers 6,000 people, 3,000 with a mask, 3,000 without.

No difference on Covid but increase in infections in the group….

masks!

"

In this Twitter post, another internet user highlights the DANMASK-19 study.

- Tom Hollmann

The authors of the study admit, however, that their results are not "statistically significant". 

 takes stock.

FAKE OFF

Danmask-19 is a study conducted by 21 Danish researchers between April and June 2020. It aims to "estimate whether wearing a surgical mask outside the home reduces the risk of Sars-Cov-2 (Covid-19) infection in a context where wearing a mask is unusual, and not recommended by public health measures ”.

In fact, wearing a mask was not recommended in Denmark over the period covered by the study.

Health authorities, on the other hand, advocated social distancing and the quarantine of suspected cases, and bars and restaurants closed until May 18.

For the Danmask study, 6,000 people were selected, then tested for Covid-19.

The negative people were separated into two groups, one wearing the mask and the other not wearing it.

All participants were then urged to spend at least three hours a day outside their homes for a month, before undergoing a new battery of tests (serological and PCR).

A first group carried out the experiment from mid-April to mid-May, the other throughout the month of May.

In the end, the study concludes that 1.8% of mask wearers contracted Covid-19, compared with 2.1% of those who did not wear it.

First point: contrary to what some Internet users claim, there is no more infection in the group of masked people.

In view of these results, you might think, like Florian Philippot, that wearing a mask outside the home is useless.

However, one of the authors of the study, Kasper Iversen, of the University of Copenhagen, himself declared that the current recommendations on wearing a mask "are not seriously questioned by the study" .

A study with too much bias

To better understand Danmask-19,

called on Didier Lepelletier, co-chair of the permanent working group on Covid-19 of the High Council for Public Health and professor of hospital hygiene at the Nantes University Hospital.

He is also a “reviewer” for more than thirty scientific journals.

"It is an interesting clinical study, published in a recognized journal, but which has too much bias to provide significant results," says the professor.

First, and as its authors indicate, the study is not "randomized", neither single nor double blind.

This means that each of the parties involved in the clinical study knows the result of the drawing of lots carried out to separate the study groups.

Single blind, it is the patient who does not know what has been assigned to him (a drug or a placebo, as part of a pharmaceutical study, for example).

Double-blind, this is also the case for researchers.

The advantage of this type of protocol is to reduce the influence that knowledge of the information could have on the person analyzed, as well as on the researcher.

"Here, for example, it would have been necessary to distribute both effective masks and masks which are less effective", notes the professor.

A particular context

Then, it is the general conditions of the study that make the results difficult to interpret, for Professor Didier Lepelletier.

The study group tested from mid-April to mid-May, for example, carried out the experiment in full containment and in a country with an extremely low transmission rate, which greatly limits the risk of contamination.

"Travel outside the home was also not controlled, and there is nothing to indicate that people infected with Covid-19 could not be once at home, without masks", adds. he.

Participants were also not trained to wear masks.

If they nevertheless received a user manual, the authors thus indicate that 46% of the participants “followed the recommendations” on wearing the mask, that 47% “generally respected” them, and that 7% did not. did not follow them.

A statistical interval that is far too wide

Finally, the most important bias is undoubtedly statistical.

In the results highlighted on social networks, detractors of the mask focus on the 23% increase in the risk of infection for the group of masked people.

But the sentence published in the journal 

taken as a whole: "Although the results are not statistically significant, the 95% confidence interval (CI) is compatible with a reduction of 46% to a 23% increase in risk of infection, ”say the co-authors.

“It's a far too wide interval!” Explains Didier Lepelletier.

I'll spare you the course in biostatistics, but neither the confidence interval (CI), nor the odd ratio (OC) allow us to arrive at conclusive results.

To say that the best-case mask reduces the risk of infection by 46% and that in the worst-case scenario increases it by 23% is not sufficient to draw conclusions from this study, and the authors readily admit it.

"

"No data on the effectiveness of wearing a general mask"

The authors of the study warn about the rapid interpretations that could be made of their work: “These results do not provide data on the effectiveness of generalized mask wearing in [a community] to reduce Sars infections. -Cov-2 […].

The results also suggest that people should not abandon other security measures against Covid-19, regardless of the use of masks.

"

An opinion shared by Didier Lepelletier who, in accordance with the report of the High Council for Public Health, indicates "that in the general population, wearing a mask, including by asymptomatic people, greatly reduces the transmission of Sars-Cov-2, in conjunction with other barrier measures.

"

Media

Report information that you think is false to the "Fake Off" team of "20 Minutes"

Science

Coronavirus: How to properly read a scientific study?

• Covid 19

• Medicine

• Fake off

• Science

• study

• Fact checking

• Society

• Coronavirus